We suggest that increased capacity for genetic instability is fundamentally associated with malignant progression of breast cancer and results in cells which have gained the capacity to grow at metastatic sites and/or to adapt to chemotherapeutic drugs by readily generating resistant clones. Therefore, sensitive assays of genetic instability might predict which tumors will be likely to recur early and/or which might respond poorly to drug therapy. In this project, we propose to evaluate breast cancers for two aspects of genetic instability, the first class being loss of allelic hetereozygosity measured by loss of restriction fragment length polymorphism (RFLP). There are two rationales for this approach. First, random deletions throughout the chromosomes may be sensitive indicators of generalized chromosome instability, thereby indicating those tumors with poor prognosis. Secondly, it is possible that we may find specific deletions that are relevant for human breast cancers by examining those chromosomes for which gross chromosomal abnormalities have been nonrandomly associated with breast cancer. If deletion of a given locus is important for malignancy, one might expect that some tumors would have gross abnormalities of that chromosome while others might have only small deletions that could be detected only by a more sensitive assay such as RFLP. Our second approach will be to examine breast cancers in short-term culture for their capacity to amplify genetic material in response to selective pressure. Specifically, we will measure capacity to amplify dihydrofolate reductase in response to methotrexate, an inhibitor of this enzyme.